Biodegradable polymeric scaffolds that can be injected into a bone defect and crosslinked in situ may provide surgeons with minimally invasive treatment options for several clinical situations, among them segmental skeletal defects and posterior spinal fusion. Scaffold design parameters include biocompatibility, rheologic propertries that detrermine injectability, degradation rate and end products, mechanical properties, and osteoinductivity. We hypothesize that 1) interpenetrating polymer networks (IPN) formed by inter-crosslinking of poly(propylene fumarate) (PPF), which contains unsaturated double bonds for in situ crosslinking, and a novel poly(caprolactone fumarate) (PCLF), which contains flexible poly(caprolactone) macromers in its backbone, may self-crosslink in the absence of a crosslinking agent, 2) the use of degradable hydrogel microspheres as the porosity generating method would both improve the scaffold's injectability and eliminate the step of solid porogen diffusion from the scaffold needed in leaching techniques for pore generation, and 3) controlled delivery of bone morphogenic protein (BMP-2) from the scaffold at the bone regeneration site could be optimized to take maximum advantage of this molecule's osteoinductive properties. Therefore, we propose to address the issues of self-crossllnking, degradation, injectability, and controlled delivery of growth factors. In the first aim, the self-crosslinking and degradation characteristics of a novel PPF/PCLF IPN will be investigated to eliminate the use a crosslinking agent in injectable systems. In this aim, we will also sytnthesize natural and synthetic biodegradable hydrogels for use as porogens to improve the scaffold's rheological properties during injection. In the second aim, BMP-2 will be encapsulated in biodegradable PPF/PLGA blend microspheres, the microspheres will be immobilized in the IPN/hydrogel composite scaffold, and we will study the release kinetics of the growth factor as well as its bioactivity. In the third aim, the composition of the hydrogel as the porogen will be optimized by measuring the in vitro migration rate of marrow stromal cells (MSC) through candidate porogens. The function of MSCs in vitro will be used to optimize the loading dose of BMP-2 microspheres in the composite scaffold. This will occur by measuring the MSC's proliferation and secretion of both alkaline phosphatase and osteocalcin in response to BMP released from the scaffold. In the fourth aim, a rat segmental defect model and a rabbit posterolateral spine fusion model will be used to assess the ability of the IPN/hydrogel composite scaffold with loaded BMP-2 microspheres to induce bone formation in vivo.